The Circadian Gene NPAS2, a Putative Tumor Suppressor, Is Involved in DNA Damage Response Free

Mounting evidence that the circadian clock may function as a tumor suppressor at the systemic, cellular, and molecular levels through its involvement in cell proliferation, apoptosis, cell cycle control, and DNA damage response increasingly suggests a novel role for circadian rhythm and circadian genes in tumorigenesis (1, 2). Our recent genetic association studies have shown that a missense mutation (Ala394Thr) in neuronal PAS domain protein 2 (NPAS2), a core circadian gene and transcriptional regulator (3, 4), is a biomarker for individuals' susceptibility to two human cancer types, breast cancer and non–Hodgkin's lymphoma (5, 6). Although these molecular epidemiologic findings are the first evidence that NPAS2 might be involved in tumorigenesis, little is known about the underlying mechanisms.

NPAS2, the largest circadian gene, encodes for a member of the basic helix-loop-helix-PAS class of transcription factors that forms heterodimers with BMAL1 and is involved in the regulation of biological rhythms (7, 8). The NPAS2:BMAL1 heterodimer is part of the positive circadian feedback loop and may bind to E-box sequences in the promoter regions of target circadian genes, thereby activating their expression and regulating circadian rhythm (9, 10). In addition to regulating other circadian genes, NPAS2:BMAL1 heterodimers have been shown to affect expression of the proto-oncogene c-MYC (11). This finding implies potential involvement of NPAS2 in tumorigenesis that has not been fully explored.

In order to investigate the functional role of NPAS2 in cancer-related biological pathways, we did RNA interference–mediated knockdown of NPAS2 in vitro followed by a variety of functional assays related to DNA damage response. The findings, although preliminary, will further our understanding of the role of NPAS2 in tumorigenesis, and provide an important first step in elucidating the mechanisms by which this gene may be involved in cancer-related processes. In addition, these findings will aid in directing future studies concerning the effect of the circadian system on carcinogenesis.

It is now clear that molecular clocks reside not only in a central pacemaker, but also in peripheral tissues, even in immortalized cells (12, 13). Previous evidence suggests that circadian genes, including NPAS2, show detectable circadian mRNA expression in all tissues except testis (14). Before beginning our study we confirmed the presence of measurable levels of NPAS2 expression in MCF-7 and HCT-15 cells using quantitative reverse transcription PCR, as well as the presence of NPAS2 protein in MCF-7 cells by Western blot (data not shown). We then transfected the cells with small interfering RNA (siRNA) oligos targeting NPAS2 along with a scrambled negative siRNA oligo with no known homology to the human genome (mock siRNA). NPAS2 knockdown relative to mock-treated cells was quantitatively determined by quantitative reverse transcription PCR prior to each assay, and only those populations with >70% reduction were used in subsequent analyses.

We did a pathway-based PCR expression array (SuperArray Bioscience) to determine the effect of NPAS2 silencing in MCF-7 cells on the expression profile of 84 genes with established roles in DNA damage signaling. Consistent with NPAS2's role as a transcriptional enhancer, 74 genes (91.4%) were down-regulated, and 22 genes (27.2%) were down-regulated by 2-fold or more compared with cells with normal NPAS2 levels. Only seven genes showed increased expression (8.6%), none of which were significantly different from the negative control (Fig. 1). We compared gene expression in NPAS2 knockdown cells with that of cells with normal NPAS2 expression using the Student's t test. At the nominal P value of 0.05, 27 genes were significantly down-regulated, whereas only 4.2 (84 × 0.05) would be expected to be significant by chance alone. However, the magnitude of many of these changes was small, and therefore, may not be biologically relevant. As such, we considered a gene to have significantly altered expression only if it displayed >2-fold up-regulation or down-regulation, and had a P value of <0.05. Seven genes with a variety of functional roles fit this criteria for significant down-regulation, including genes important for cell cycle checkpoint control (FANCG), damaged DNA binding (DMC1, DDB1, and MSH2), double-strand break repair (PRKDC), cell cycle arrest (MAPK12), and mismatch repair (EXO1; Table 1). In order to further confirm these findings, and to determine whether the observed response was specific to MCF-7 cells, we repeated the NPAS2 knockdown experiment and PCR expression array using a colorectal adenocarcinoma cell line (HCT-15). Using the nominal P value of 0.05, eight genes were significantly down-regulated, and again, no genes were significantly up-regulated. Although no genes fit the original significance criteria of fold change of >2 and P value of <0.05, the three genes most significantly altered according to the t test alone were also identified as significant in MCF-7 cells. Given that none of the DNA damage–responsive genes included in the array were altered by 2-fold or more following NPAS2 knockdown in the HCT-15 cells, we did the subsequent functional analyses on MCF-7 cells only.

We did flow cytometric analyses to determine the effect of NPAS2 silencing on cell cycle progression in response to exposure to the chemical mutagen methyl methanesulfonate (MMS). MCF-7 cells were transfected with either NPAS2-targeting or scrambled negative siRNA. Forty-eight hours after transfection, cells were treated with either 0.015% of MMS or PBS for 1 hour, washed, and then left under normal conditions for an additional 24 hours. Harvested cells were then analyzed by flow cytometry and cell phases were determined. As expected, the cell phase distribution of the cell population with normal NPAS2 expression was significantly different among those treated with PBS compared with MMS (χ²₂ = 13.777; P = 0.001; Fig. 2). Mutagen-treated cells with normal NPAS2 expression were more likely to be in G₁ or G₂ phase, corresponding to the major cell cycle checkpoints for DNA damage repair, whereas the number of cells in S phase was significantly reduced. However, among cells with reduced NPAS2 expression, there was no significant difference between cells treated with PBS and those treated with MMS (χ²₂ = 3.830; P = 0.147), indicating an aberrant response to DNA damage. Furthermore, there was a significant difference in the cell cycle distribution of normal and NPAS2 knockdown populations among cells not treated with mutagen (χ²₂ = 6.096; P = 0.0475). This difference was even more pronounced among mutagen-treated cells (χ²₂ = 10.109; P = 0.006).

The alkaline single-cell gel electrophoresis assay (comet assay) was used to assess differences in DNA damage accumulation and repair capacity between cells with normal NPAS2 expression and those with reduced NPAS2 expression in MCF-7 cells. After the siRNA and MMS/PBS treatments mentioned above, cells were washed in PBS, then incubated in normal medium for 3 hours, followed by comet assay analysis. The mean Olive tail moment was determined for 50 cells from each of the four treatment groups (siRNA-positive/siRNA-negative and MMS-positive/MMS-negative) and all pairwise comparisons were examined using Fisher's protected least significant difference (Fig. 3). Longer “tails” associated with an individual cell corresponded to greater accumulation of damage, as the head consists of intact DNA whereas the tail material is comprised of damaged DNA. Among the samples not treated with mutagen, the difference in tail moment between cells with reduced and normal NPAS2 expression was not significant (P = 0.668), indicating that without inducing damage, there is no difference in repair. However, among the mutagen-treated populations, cells with reduced NPAS2 had significantly longer mean tail moments (P = 0.002), suggesting decreased DNA repair capacity compared with cells with normal NPAS2 expression. All other comparisons were significant, as expected.

A similar protocol was used to assess differences in cell viability and apoptotic response to mutagen exposure between cells with normal NPAS2 expression and those with reduced levels of NPAS2. MCF-7 cells were treated as described in the cell cycle experiment, with the exception that cells were harvested 48 hours after MMS or PBS treatment to allow additional time for the initiation of the apoptotic response. Percentages were then determined for early apoptotic, live, and dead cells for each treatment population. For the cell viability comparison, siRNA-positive samples were compared with siRNA-negative samples for each MMS or PBS treatment. For the apoptosis comparison, MMS-treated cells were compared with PBS-treated cells in each of the siRNA treatment groups. In both mutagen-treated and nontreated populations, cells with reduced NPAS2 expression were significantly less viable than normal cells (P = 0.001 and P = 0.003, respectively), indicating that NPAS2 knockdown alone was sufficient to elicit increased vulnerability to cell death (Fig. 4). However, in both NPAS2 knockdown and normal cells, treatment with MMS resulted in significantly increased levels of apoptosis (P = 0.037 and P = 0.027, respectively), suggesting that NPAS2 may not directly hinder the apoptotic process.

It has been suggested that the disruption of circadian rhythm may have broad consequences at the molecular level, as previous evidence has shown that 2% to 10% of all mammalian genes display circadian oscillation (15, 16). The results presented here contribute to increasing evidence suggesting an important role for NPAS2 in transcriptional regulation, but the full extent to which it is involved in disease-related pathways remains to be elucidated (4). Although previous reports have suggested that NPAS2 may directly regulate other genes in the circadian regulatory system (9, 10), our current study shows that genes involved in cancer-related biological pathways are also influenced by NPAS2, which could explain the observed functional effect of NPAS2 knockdown on cell cycle progression and DNA repair capacity, as well as the epidemiologic evidence suggesting an association between polymorphisms in NPAS2 and cancer risk (5, 6). In addition to previous studies demonstrating the importance of another circadian gene, PER2, in tumor suppression (17, 18), our findings further highlight the relevance of the circadian system to carcinogenesis, pointing to an area that is perhaps understudied in the field of cancer research (19).

Although it is not clear whether any of the genes identified in the PCR expression array are regulated by NPAS2 directly, it is of note that of the eight functional gene groups represented in the array (apoptosis, cell cycle arrest, cell cycle checkpoint, double-strand break repair, damaged DNA binding, base-excision repair, mismatch repair, and other genes related to DNA repair) only the apoptotic pathway did not have a single gene significantly affected by NPAS2 knockdown in either cell type. This is consistent with the functional findings which suggest that NPAS2 silencing decreases cell viability, results in aberrant cell cycle response to DNA damage, and decreases DNA damage repair capacity, but does not alter the apoptotic response. It is also of note that the array results imply a significantly more prominent role for NPAS2 in breast cancer cell lines compared with cells derived from colorectal cancer. Although expression of some genes in the array was affected by NPAS2 knockdown in the HCT-15 cells, the effect was much less profound than that observed in MCF-7 cells, and none reached our significance criteria of P < 0.05 and >2-fold expression change. These data suggest that the role of NPAS2 may vary according to cancer type, and as such, the results presented here, although relevant to breast cancer, may not necessarily translate uniformly across other cancer types.

An important limitation of our methodology is that cells were only transiently transfected with siRNA oligos, and therefore, silencing of NPAS2 could not be stably induced and successfully transfected cells could not be exclusively selected. As a result, we could not assess the long-term effects of reduced NPAS2 expression on cell viability or other relevant pathways, nor could we achieve 100% silencing of NPAS2. However, because the reduction in NPAS2 expression was quantified prior to each assay, this does not affect the interpretation of the results presented here, but it does provide an avenue for further research. In addition, these findings provide only indirect support for the role of NPAS2 as a transcriptional regulator of cancer-related genes. More direct evidence is needed in order to determine whether the genes implicated here are directly targeted by NPAS2, or whether their expression is altered indirectly via some other mechanism. This could be achieved via ChIP-on-chip assay using NPAS2-specific antibodies and a targeted or whole-genome promoter array.

Together, our observations indicate important functional consequences of the circadian gene NPAS2 on DNA damage response, a key pathway related to tumorigenesis. Although epidemiologic evidence has linked NPAS2 with a variety of disorders, including winter depression (20) and autistic disorder (21), as well as cancer (5, 6), this is the first functional evidence demonstrating a broad role for NPAS2 in tumorigenesis. Although this study does not definitively address all of the questions relevant to the relationship between NPAS2 and carcinogenesis, we do believe that it is an important first step in elucidating the mechanisms by which this gene may be involved in tumorigenesis. These findings lend further support to the view that circadian genes may serve as a novel group of tumor suppressor genes and provide a molecular basis for the epidemiologic observations that disruption of circadian rhythm is a risk factor for the development of several cancer types (22-24).

Human breast adenocarcinoma cells (MCF-7) and colorectal adenocarcinoma cells (HCT-15) were obtained from American Type Culture Collection. MCF-7 cells were maintained in DMEM (Invitrogen) supplemented with 10% fetal bovine serum (Invitrogen) and 0.01 mg/mL of bovine insulin (Sigma-Aldrich). HCT-15 cells were maintained in RPMI 1640 (Invitrogen) supplemented with 10% fetal bovine serum. For comparisons between cells treated with scrambled negative and NPAS2-targeting siRNA oligos, cells for each treatment were harvested from the same colony at the same time. In addition, all experiments were run in parallel to control for the effects of time and cyclic gene expression.

siRNA oligos targeting NPAS2 (duplex 1: sense, GGGUGCCUCGGGCCUUAAUACCAGT; antisense, ACUGGUAUUAAGGCCCGAGGCACCCAC; duplex 2: sense, CCCAGGGUCCAAAGCCAAUGAGAAG; antisense, CUUCUCAUUGGCUUUG-GACCCUGGGUU; duplex 3: sense, GGCACACUCUAUAGCCAUACUGGAC; antisense, GUCCAGUAUGGCUAUAGAGUGUGCCCA), a scrambled-sequence negative control siRNA with no known homology to the human genome to control for the effects of transfection, a positive control targeting the constitutively expressed housekeeping gene HPRT, and a fluorescent-labeled transfection efficiency control were chemically synthesized by Integrated DNA Technologies. Each was diluted in Opti-Mem I reduced serum medium (Invitrogen) and complexed with Lipofectamine 2000 transfection reagent (Invitrogen) prior to transfection according to the manufacturer's instructions. Each oligo was added to a 12-well culture dish followed by ∼100,000 cells suspended in normal growth medium, for a final siRNA concentration of 10 nmol/L (reverse transfection). After 48 h, cells were either harvested for quantitative real-time reverse transcription PCR, or incubated with either PBS (pH 7.4; negative control) or 0.015% (v/v) of MMS (mutagenic agent) for use in subsequent assays. Cells transfected with the fluorescent oligo were visualized by fluorescence microscopy using an FITC filter. Fifty randomly selected cells were assessed for oligo uptake, and >90% were found to be successfully transfected.

RNA was isolated from both MCF-7 and HCT-15 using the RNA Mini Kit (Qiagen), with on-column DNA digestion, according to the manufacturer's instructions for mammalian cells. Quantitative real-time reverse transcription PCR conditions were prepared using the Quantitect one-step kit (Qiagen) according to the manufacturer's protocol. Primers for 18s rRNA and NPAS2 were designed in-house and chemically synthesized by IDT (18s, 5′-ATCAACTTTCGATGGTAGTCG-3′ and 5′-TCCTTGGATGTGGTAGCCG-3′; NPAS2, 5′-TCTGGATCACAGAGCACCTC-3′ and 5′-CAGGAGCTCCAGGTCATCA-3′). Standard duplicated quantitative reverse transcription PCRs with disassociation curves were done on a Stratagene MX3000P instrument (Stratagene). RNA quantity was normalized to 18s rRNA, and changes in RNA expression were quantified according to the 2^−ΔΔCt method. To detect NPAS2 protein expression in MCF-7, we also did a standard Western blot using lysate of MCF-7 and rabbit polyclonal anti-NPAS2 (sc-28708; Santa Cruz Biotechnology).

Once NPAS2 knockdown was confirmed (>2-fold reduction by quantitative reverse transcription PCR), the human DNA damage signaling RT² profiler PCR array (SuperArray Bioscience) was used to assess the effect of NPAS2 reduction on the expression of 84 genes with known roles in DNA damage response. Synthesis of complementary DNA, real-time PCR, and statistical analyses were done according to the manufacturer's instructions, and the data shown represent the average of two replicates. Cycle thresholds were determined for each PCR product, with amplification curves occurring later than a cycle 35 considered failed or unexpressed. Five housekeeping genes were included as RNA content controls, and the three with the lowest standard deviations across replicates were used in the analysis (for MCF-7: B2M, GAPDH, and ACTB; for HCT-15: B2M, GAPDH, and HPRT1). The average difference in cycle threshold between the first and second replicates for these genes in the siRNA-positive and siRNA-negative samples was 0.0 and 0.1 cycles for MCF-7, and 0.0 and 0.4 for HCT-15, respectively, indicating excellent stability and reproducibility. In addition, a genomic DNA control primer set was included in each PCR run which can detect the presence of DNA contamination in each sample. The Ct for this product is compared with the average of the housekeeping genes, with a ΔCt value of <4 indicating contamination. No sample came close to reaching this cutoff, and the average across all samples was 9.5 cycles (range, 7.5-10.3) for MCF-7. No genomic DNA was detected in the HCT-15 samples (Ct > 35). All samples also passed additional quality controls for reverse transcription and PCR efficiency provided by the manufacturer. Finally, three genes with failed PCR amplifications in one or more replicates were removed from the MCF-7 analysis.

For cell cycle analysis, cells were transfected with siRNA duplex 1 or negative control and incubated under normal conditions for 48 h (37°C, 5% CO₂). Cells were then treated for 1 h with either MMS or PBS, washed and left overnight, and then harvested for analysis. Cells were incubated with propidium iodide stain and analyzed by flow cytometry using a fluorescence-activated cell sorter flow cytometer (Becton Dickinson) according to methods previously described. More than 40,000 cells were counted, and data was analyzed using the FlowJo flow cytometry analysis software (Tree Star, Inc.). Cell phases were determined by the software using the Watson pragmatic algorithm. For apoptosis and cell viability analyses, cells were stained with propidium iodide along with Annexin V using reagents in the Vybrant Apoptosis assay kit no. 2 (Invitrogen), according to the manufacturer's protocol. Ten thousand cells were counted, and each was categorized as live, dead, apoptotic, or ambiguous using the FlowJo flow cytometry analysis software (Tree Star, Inc). Cells that could not be definitively called by the software were excluded from the analysis, and percentages were calculated using the remaining population. All data are the result of two independent experiments.

DNA damage was assessed using the alkaline single-cell gel electrophoresis assay (comet assay). After the siRNA and MMS/PBS treatments mentioned above, cells were washed in PBS, then incubated in normal medium for 3 h. Cells were then harvested and mixed with low-melting agarose and fixed onto slides. The cells were then lysed and DNA was denatured using an alkaline (pH > 14) solution at 4°C. Slides were then placed on a horizontal electrophoresis device which allowed damaged and/or broken DNA to migrate away from the nucleus. Slides were then stained with ethidium bromide, visualized by fluorescence microscopy, and analyzed using the Komet 5 comet assay analysis software which quantitatively determines the extent of DNA damage in each sample using the mean Olive tail moment calculation as previously described (25).

All statistical analyses were done using the SAS statistical software (SAS Institute), unless otherwise noted. Differences in cell cycle progression for cells treated with scrambled negative siRNA and NPAS2-targeting siRNA were determined using the Pearson χ² test comparing the percentages of the total cell population in each phase of the cell cycle for both cell populations. Differences in apoptotic response were determined by the Student's t test comparing the percentage of early apoptotic cells in each cell population. Cell viability was also determined by t test of the percentage of live cells, and in both instances, tests for equality of variance were not significant and pooled variances were used. Data was collected on 10,000 cells per sample and the data shown are based on duplicate experiments. Statistical analysis of DNA damage repair capacity was done using the Komet 5 comet assay analysis software. Differences in DNA repair were determined for all pairwise associations of the four treatment groups (siRNA-positive/siRNA-negative and MMS-positive/MMS-negative) using Fisher's protected least significant difference for the mean Olive tail moment. Gene expression differences in the PCR expression array were determined using the 2^−ΔΔCt method. If the gene was down-regulated, fold expression change was defined as −1 / (fold difference). For example, if the fold difference (2^−ΔΔCt) of a gene was 0.5, this would be expressed as a 2-fold down-regulation, or a fold change of −2.

No potential conflicts of interest were disclosed.

We are grateful to Chunhui Yi for assistance with antibody screening and Western blot. We also thank the flow cytometry core at Yale University for assistance with cell cycle and apoptosis assays.